Nanoparticles of Cu–Co alloy derived from layered double hydroxides and their catalytic performance for higher alcohol synthesis from syngas

Abstract

A series of layered double hydroxides (LDHs) with different Cu/Co ratios were prepared according to the co-precipitation method and used as catalyst precursors for higher alcohol synthesis. The prepared samples were characterized by XRD, TPR, SEM, TEM and BET techniques. After calcination, the LDHs were transformed into a mixture of CuO, Co₃O₄, CuCo₂O₄ and alumina, and these oxides were mixed uniformly with sizes of several nanometers. For the sample with a ratio of Cu/Co = 1/2, the copper and cobalt species are mainly in the CuCo₂O₄ phase. In the reduction process, the superior mixed copper and cobalt species in nanoscale were reduced to Cu–Co alloy, which was confirmed by the XRD and TEM results. The prepared bimetallic catalysts showed high activity, good stability and very high selectivity to C₂₊ alcohols. With the best catalyst, CO conversion of 51.8%, selectivity to alcohols of 45.8% and 94.3 wt% of C₂₊OH in the total alcohols were obtained at 250 °C, 3 MPa and GHSV of 3900 mL (g_cat h)⁻¹.

---

1. Introduction

With the dwindling of petroleum resources and the increasing consumption for energy, developing alternative fuels for sustainable world economics are urgently needed.^1,2 Higher alcohol synthesis (HAS) via syngas from coal, natural gas or renewable biomass has attracted a considerable amount of attention due to the potential application of higher alcohols as motor fuels, gasoline blend additives and other intermediates for chemical feedstocks.^3–5

However, the existing technologies for HAS are still in the laboratory stage, one of the critical problems is that the selectivity to C₂₊ alcohols is too low. For most of the catalysts reported, methanol is the main product instead of higher alcohols. The price of each of the C₂₊ alcohols, including ethanol, propyl alcohol and butyl alcohol, is more than twice that of methanol, and production of methanol has been industrialized for tens of years with methanol selectivity close to 100%. Therefore, only selectivity to C₂₊ alcohols is high enough, excluding separation cost, to allow HAS to be of value for industrialization.

Catalysts for HAS from syngas can be categorized into four types, including modified methanol catalysts, Rh-based catalysts, Mo-based catalysts and modified Fisher–Tropsch catalysts. Modified methanol synthesis catalysts produce mainly methanol.⁶ Rh-based catalysts possess a comparatively high selectivity to ethanol and good activity,^7,8 while the price of Rh is too high. Mo-based catalysts exhibit excellent sulfur resistance, however, they must be operated at high pressures and temperatures, and usually a long activity induction period before the reaction is required.^9,10

The modified Fisher–Tropsch catalysts include Cu–Fe and Cu–Co based catalysts, of which Cu–Fe catalysts tend to initiate the water–gas shift reaction (WGSR) and thus generate more CO₂. Comparatively, Cu–Co based catalysts show good performance for HAS at mild reaction conditions, and have become one of the most promising catalysts for practical application.

For Cu–Co based catalysts, copper acts as the active site for the non-dissociative adsorption of CO and the dissociative chemisorption of H₂; while cobalt is responsible for the dissociative activation of CO and carbon chain growth. Therefore, the synergistic working of the copper and cobalt sites can promote the generation of higher alcohols.^11–13 It is acknowledged that the formation of Cu–Co alloy is beneficial for the synergistic catalysis. Hence the formation of Cu–Co alloy is important for HAS, and the formation of Cu–Co alloy can increase the selectivity to C₂₊ alcohols.^14–17

In order to form Cu–Co alloy, CuCoO₂ (ref. 18) and CuCo₂O₄ (ref. 14 and 19) were used as the precursors, in which cobalt and copper ions are uniformly distributed at the atomic level. Consequently, the formation of Cu–Co alloy should be favored in the reduction process. However supported catalysts generally are prepared by an impregnation method followed by drying and calcination. During the impregnation process, making the metal species uniform and highly dispersed on the support is vital to the performance of catalyst. There are many influencing factors, such as the interaction between the metal ions and the support, and the migration and aggregation of the metal species during the drying and calcination processes. Thus, in the resulting supported Cu–Co based catalysts, apart from CuCoO₂ and CuCo₂O₄, CuO and Co₃O₄ would be inevitably formed. After reduction, mono-metal particles of copper and cobalt derived from CuO and Co₃O₄ would be generated, reducing the selectivity to C₂₊ alcohols by producing methanol and hydrocarbons, respectively.

Layered double hydroxides (LDHs), a class of synthetic two-dimensional nanostructured layered materials,^20–22 have recently received significant attention owing to their promising applications as heterogeneous catalysts^23–25 and catalyst supports.²⁶ LDHs can be expressed by a typical formula [M_1−x²⁺M_x³⁺(OH)₂]^x+(Aⁿ⁻)_x/nmH₂O, where M²⁺ and M³⁺ are divalent and trivalent cations, respectively. Aⁿ⁻ is the interlayer anion, to create electrical neutrality with the cations. In LDHs, the metal cations of M²⁺ and M³⁺ are uniformly distributed at an atomic level in the brucite-like layers. One character of LDHs is that the uniform distribution of the metal species is maintained in the resultant mixed metal oxides after calcination although the LDHs structure is destroyed, which is owing to the topological effect of LDHs.^27,28 The uniform distribution of the metal species in the mixed metal oxides and the topological effect of LDHs are beneficial and crucial for preparing supported nano-metal alloys.

Studies on bimetallic catalysts prepared via LDHs as the precursors can be found. For instance, He et al.²⁴ prepared Co–Ni based catalysts by calcining and reducing LDHs precursors, which exhibited high activity and good stability for ethanol steam reforming. Yang et al.²⁹ reported that uniform Co–Fe alloy nanoparticles with a narrow size distribution were obtained by reducing the LDH precursors which contained Mg, Co and Fe ions. However, to the best of our knowledge, using LDHs as the precursor to prepare supported nanoparticles of Cu–Co alloy for HAS has not been reported.

In the present investigation, Cu–Co alloy nanoparticles supported on Al₂O₃ were prepared using LDHs as the precursor. The resultant catalysts showed good catalytic performance for HAS from syngas, and especially exhibited high selectivity to C₂₊ alcohols.

2. Experimental

2.1 Materials

Analytical grade chemicals of Cu(NO₃)₂·3H₂O, Co(NO₃)₂·6H₂O, Al(NO₃)₃·9H₂O, NaOH and Na₂CO₃ were purchased from Aladdin and used without further purification. Deionized water was used in all the experimental processes.

2.2 Preparation of the Cu–Co catalysts

A series of (Cu_xCo_y)₂Al-LDHs (([Cu²⁺] + [Co²⁺])/[Al³⁺] = 2 in molar) with Cu/Co molar ratios of 1 : 0, 1 : 1, 1 : 2, 1 : 3 and 0 : 1 was prepared using a co-precipitation method. Cu(NO₃)₂·3H₂O, Co(NO₃)₂·6H₂O and Al(NO₃)₃·9H₂O were dissolved in deionized water at [Cu²⁺] + [Co²⁺] + [Al³⁺] = 1.0 M. NaOH and Na₂CO₃ were dissolved in deionized water with [NaOH] = 1.5 M and [CO₃²⁻] = 2[Al³⁺]. The two solutions were simultaneously added into deionized water under vigorous stirring, maintaining pH = 9.5 ± 0.1. As the precipitation was completed, the slurry was aged at 80 °C for 12 h, and then filtered, washed with distilled water and dried at 70 °C for 12 h. The thus prepared LDHs were calcined at 500 °C for 3 h, denoted as (Cu_xCo_y)₂Al-CLDHs, and then reduced at 450 °C for 3 h in hydrogen. The reduced catalysts were named as (Cu_xCo_y)₂Al-Red.

2.3 Catalyst characterization

The phase analyses of the catalysts were determined by the X-ray Diffraction (XRD) technique which was recorded on an X’Pert Pro instrument, a Co Kα radiation source was used with an accelerating voltage of 40 kV and electric current of 40 mA. The spectrum was collected at a scanning speed of 5° min⁻¹.

Nitrogen adsorption and desorption isotherms were performed on Tristar 3000 micromeritics apparatus at −196 °C. The specific surface areas were calculated based on the BET method and the pore size distributions were calculated from the adsorption branch of the isotherms using the BJH model. All samples were outgassed under vacuum at 300 °C for 4 h prior to analysis.

Scanning electron microscopy (SEM) characterizations were performed on a Hitachi S-4800 field-emission scanning electron microscope to observe the morphology of the samples, which were treated with Au sputtering first.

Transmission electron microscopy (TEM) pictures and high angle annular dark field scanning transmission electron microscopy (HAADF-STEM) line scans combined with energy dispersive X-ray spectroscopy (EDX) for the determination of metal composition and elemental mapping were obtained on a JEOL JEM-2100F microscope field-emission transmission electron microscope. After ultrasonic dispersion of the catalysts in absolute ethanol, the well-dispersed samples were dropped and dried on a Mo grid with a layer of holey carbon film.

Temperature programmed reduction (TPR) tests were carried out using a fixed bed micro-reactor in order to study the reducibility of the metal species in the catalysts. In each run, 50 mg of the catalyst was loaded into a quartz tube reactor and pretreated with 5% H₂/Ar to remove the air in the reactor at room temperature. Then, the catalyst was heated from room temperature to 800 °C at a heating rate of 10 °C min⁻¹ in the presence of 5% H₂/Ar flow at a flow rate of 48 mL min⁻¹.

2.4 Catalyst characterization

The higher alcohol synthesis tests were conducted in a fixed-bed, stainless flow micro-reactor with a length and inner diameter of 300 nm and 8 nm, respectively. 800 mg of the catalyst with 40–60 mesh grain size was diluted with 3200 mg quartz sand with the same size and then loaded into the reactor, the mass ratio of the catalyst and quartz sand was 1 : 4. All catalysts were activated on-line under a flow of H₂ at 450 °C for 3 h at a rate of 2 °C min⁻¹. Subsequently, the reactor was cooled to room temperature, and the pressure was increased to 3 MPa by adding the syngas mixture (H₂/CO/N₂ = 8 : 4 : 1, where N₂ was used as the internal standard gas for analyzing the reacted gases). The gas hourly space velocity was set at 3900 mL (g_cat h)⁻¹. At each reaction temperature from 250 to 290 °C, the reaction was maintained for 1.5 h to obtain the steady-state activity and selectivity; except at 240 °C the reaction was conducted for 3 h. A gas chromatograph with two packed columns was used to analyze the products. The gas products of CO, H₂, CH₄, CO₂ and N₂ were separated online using a TDX-01 packed column (2 m) connected to a TCD detector. After separation by condensation, the liquid products and hydrocarbons were analyzed off-line using a Porapak Q column (3 m) connected to a FID detector.

CO conversion (X_CO) and the product selectivity (S_i) were calculated according to the following equations:

where CO_in, and CO_out are the moles of CO in the feed-gases and off-gases, respectively; n and C_i represent the number of carbon atoms in the molecule and the moles of a carbon-containing product, respectively.

3. Results and discussion

3.1 XRD

The XRD patterns of the as-synthesized (Cu_xCo_y)₂Al-LDHs are shown in Fig. 1(I). All samples show sharp and symmetric diffractions at 2θ values of 13.5, 27.2 and 40.5°, ascribed to the crystalline planes of (003), (006) and (009) for LDHs, respectively, indicating well-crystallized hydrotalcite structures. Not any detectable impurity phases can be observed, indicating the high purity of the LDH structure in the samples.

Fig. 1 XRD patterns of (I) (Cu_xCo_y)₂Al-LDHs, (II) (Cu_xCo_y)₂Al-CLDHs and (III) (Cu_xCo_y)₂Al-Red with x/y ratios of (a) 1 : 0, (b) 1 : 1, (c) 1 : 2, (d) 1 : 3 and (e) 0 : 1. Crystalline phases: (●) Co₃O₄/CuCo₂O₄, (♥) CuO, (♣) Co, (♠) Cu, (♦) Cu–Co alloy; (III′) the variation of Cu/Co molar ratio vs. lattice parameter a.

Particularly, the XRD patterns show a variation trend correlated to the chemical compositions. With an increase in the content of copper, the diffraction peaks move to slightly higher 2θ values (Fig. 1(I) inset). This is attributed to the distortion of copper ions in LDHs as pointed out by Khan et al.³⁰ Similarly, Liu et al. also pointed out that Cu²⁺ ions in LDHs show a Jahn–Teller effect that favors the formation of a distorted octahedral structure,³¹ which leads to the shift of the diffraction peaks. Compared with copper ions, cobalt ions more easily form the LDH structure, thus the two diffraction peaks of (110) and (113) at 2θ values of around 71 and 74° become sharper and stronger with cobalt content increasing.

The basic spacing (d₀₀₃), varies between 0.746 and 0.752, which is close to the values for CO₃²⁻ containing LDHs with a M²⁺/Al³⁺ ratio of about 2.³² Thus, CO₃²⁻ was successfully introduced into the space between the hydrotalcite laminates.

Fig. 1(II) shows the XRD patterns of (Cu_xCo_y)₂Al-CLDHs. It can be seen that the diffraction peaks corresponding to LDHs disappear, and the diffraction peaks of metal oxides appear, suggesting that the LDH structures are completely destroyed and transformed into metal oxides after calcination. In addition, the diffraction peaks corresponding to Al₂O₃ cannot be observed, indicating that Al₂O₃ is in an amorphous state.

For Cu₂Al-CLDHs, only a peak attributed to the CuO crystalline phase at 2θ = 41.4° exists, and the peak is broad, meaning the size of the CuO nanoparticles is small.

As for the cobalt containing samples, the diffraction peaks corresponding to the CuO phase can hardly be seen, which is attributed to the following two reasons. One is that the copper and cobalt species are combined to form CuCo₂O₄ during calcination, which is supported by the presence of the characteristic diffraction peaks of CuCo₂O₄. The other is that the copper species left in the CuO phase are highly dispersed or in a small quantity. As the atomic ratio of Cu/Co is equal to 0.5, which is the stoichiometric ratio of Cu/Co in CuCo₂O₄, copper should be present mainly as CuCo₂O₄. As cobalt content increases further, CuCo₂O₄ and Co₃O₄ are the major phases in the catalyst. This is revealed by the slight shift of the lattice parameter from 8.105 Å of pure CuCo₂O₄ to 8.101 Å. Both CuCo₂O₄ and Co₃O₄ belong to a spinel structure, while the lattice parameter of Co₃O₄ is slightly smaller due to the smaller size of the cobalt ions. Finally, Co₃O₄ becomes the only oxide with spinel phase for the Co₂Al-CLDHs catalyst.

In summary, copper and cobalt exist in CuCo₂O₄ accompanied with CuO or Co₃O₄ in the calcined catalysts, and copper and cobalt exist mainly in CuCo₂O₄ for (Cu₁Co₂)₂Al-CLDHs.

In addition, all of the diffraction peaks are broad, suggesting that the crystal grains are small. The metal oxides in the calcined samples came from LDHs, where the metal ions were uniformly mixed at atomic level, so the metal oxides should be uniformly mixed. In other words, CuCo₂O₄, CuO and Co₃O₄ were well mixed and highly dispersed in the amorphous Al₂O₃ matrix. The nanoparticles of the oxides interact, which further limits the migration and aggregation of those oxides. This means that the resultant mixed oxide possesses high anti-sintering ability, which agrees with the results in the literature,³³ where a ZnO/ZnAl₂O₄ nanocomposite was made using LDHs as the precursor and exhibited high thermal stability.

Fig. 1(III) shows the XRD results of the reduced catalysts. For Co₂Al-Red and Cu₂Al-Red, diffraction peaks corresponding to metal cobalt and copper can be seen clearly. No diffraction peak related to alumina can be detected, showing that alumina is in an amorphous state.

For (Cu_xCo_y)₂Al-Red containing both copper and cobalt elements, diffraction peaks corresponding to copper and cobalt containing oxides disappear, while peaks between the diffraction peaks of pure metal Cu and Co are observed, which can be clearly seen from the inset picture in Fig. 1(III). Su et al.³⁴ attributed the diffraction peaks between pure metal Co and Cu to bimetallic Co–Cu alloy. Volkova et al.¹⁸ demonstrated that the slight increase in lattice spacing for copper and a small decrease in lattice spacing for cobalt are the characteristics of Co–Cu alloy. Hence, Cu–Co alloy was formed in (Cu_xCo_y)₂Al-Red.

As shown in Fig. 1(III′), the metal lattice parameter a for (Cu_xCo_y)₂Al-Red is smaller than that of Cu⁰ (3.613 Å), larger than that of Co⁰ (3.548 Å), and decreases linearly with the increase of cobalt content. This is in accordance with the fact that the size of the cobalt atom is smaller than that of copper, supporting the formation of Cu–Co alloy.^35–37 Vegard’s law³⁵ pointed out that independent of other factors, the microscopic crystal structures of alloys depend on the atomic sizes and relative concentrations of the constituents. For Cu–Co alloy, the lattice parameter a decreases linearly with increasing Co content due to the smaller size of cobalt than that of copper, and vice versa, the linear variation is an indication of the Cu–Co alloy formation.

Although some CuO and Co₃O₄ were formed inevitably after the calcination of LDHs, considering that the metal ions are uniformly distributed at atomic level in the layers of LDHs, all the CuO, Co₃O₄ and CuCo₂O₄ with small and uniform particle sizes were mixed evenly in the amorphous Al₂O₃ matrix. Thus the interaction between them is enhanced, and consequently, Cu–Co alloy rather than separate copper and cobalt is formed after reduction.

3.2 TPR

The TPR profiles of the calcined catalyst are shown in Fig. 2. For Co₂Al-CLDHs, two reduction peaks in the temperature range of 310–410 °C and 600–780 °C are observed, according to previous reports,^14,38,39 which are attributed to the reduction of Co₃O₄ to CoO and CoO to metallic cobalt, respectively. The H₂ consumption ratio of the two peaks is close to 1 : 3, which supports the attribution. This result also suggests that cobalt species are mainly in the Co₃O₄ state, which is consistent with the XRD results.

Fig. 2 TPR patterns of the catalysts (a) Cu₂Al-CLDHs, (b) (Cu₁Co₁)₂Al-CLDHs, (c) (Cu₁Co₂)₂Al-CLDHs, (d) (Cu₁Co₃)₂Al-CLDHs and (e) Co₂Al-CLDHs.

A. Y. Khodakov et al.⁴⁰ stated that for the interaction between the metal oxide and support, the smaller the particle is, the stronger the interaction, and thus the metal oxide with a smaller size is harder to be reduced. Compared with that of Co₃O₄ reported by Chu et al.,⁴¹ the reduction temperature of Co₃O₄ in Co₂Al-CLDHs is higher, which is likely due to the smaller size of Co₃O₄ in Co₂Al-CLDHs.

For Cu₂Al-CLDHs, two reduction peaks centered at around 220 and 300 °C can be observed, corresponding to the reduction of CuO to metallic Cu, as reported in the literature.⁴² The peak at a lower temperature can be attributed to the reduction of highly dispersed CuO and the other is ascribed to CuO which is strongly interacting with the Al₂O₃ support.

For (Cu_xCo_y)₂Al-CLDHs, the reduction peaks at about 450 °C belong to the reduction of cobalt ions. The reasons are as follows. Firstly, copper ions could be reduced at a much lower temperature, seeing the reduction profile for Cu₂Al-CLDHs. Secondly, the area of this reduction peak increases with the content increase of cobalt in the catalyst. As compared to Co₂Al-CLDHs, the reduction temperature of this peak for cobalt oxide is obviously lower, which is ascribed to the formation of bimetallic Cu–Co at the lower reduction temperatures. It is known that metal particles can activate hydrogen and catalyze the reduction of metal ions nearby.^43,44

For (Cu_xCo_y)₂Al-CLDHs, the reduction peaks in the temperature range of 150 to 250 °C correspond to the formation of bimetallic Cu–Co. Compared to the reduction profiles of Co₂Al-CLDHs and Cu₂Al-CLDHs (Fig. 2a and e), the reduction temperatures of (Cu_xCo_y)₂Al-CLDHs (Fig. 2b–d) are obviously lower. The difference between them is that (Cu_xCo_y)₂Al-CLDHs contain CuCo₂O₄, and CuCo₂O₄ is the main phase for (Cu_xCo_y)₂Al-CLDHs as the above XRD results show. Thus, the reduction peaks in the temperature range of 150 to 250 °C are attributed to the reduction of CuCo₂O₄ to metal copper and metal cobalt.

The XRD results indicate that, besides CuCo₂O₄, there are Co₃O₄ and CuO in (Cu_xCo_y)₂Al-CLDHs. The reduction of the CuO can be included in the reduction peaks from 150 to 250 °C, because the bimetal of Cu–Co reduced from CuCo₂O₄ can catalyze the reduction of copper ions, resulting in the decrease of the reduction temperature for CuO. The Co₃O₄ may be partly reduced in the temperature range of 150 to 250 °C for the cobalt species weakly interacting with alumina, and partly reduced at about 450 °C for the cobalt species interacting with alumina.

To summarise the reduction profiles of (Cu_xCo_y)₂Al-CLDHs, copper and cobalt ions are reduced to a metallic state simultaneously at 150 to 250 °C. The XRD results and the properties of the LDH precursor suggest that the copper and cobalt species in (Cu_xCo_y)₂Al-CLDHs are uniformly and highly dispersed with a small size (see Table 1). The simultaneously generated metal copper and metal cobalt grains, which are of small size and uniformly mixed, are inclined to form Cu–Co alloy.

Table 1 Physical properties and crystal sizes of the catalysts

Samples	BET surface area (m^2 g^−1)	BJH pore volume (cm^3 g^−1)	Pore diameter (nm)	Crystal size^a (nm)	Crystal size^b (nm)
a Crystal sizes of the catalysts after calcination were calculated by X-ray diffraction with the Scherrer equation.b Crystal sizes of the catalysts after reduction were calculated by X-ray diffraction with the Scherrer equation.c Data for the catalyst after a 200 hour stability test.
Co2Al-CLDHs	137	0.47	3.03	23.2	17.8
(Cu1Co3)2Al-CLDHs	123	0.45	3.14	7.2	4.3
(Cu1Co2)2Al-CLDHs	106	0.39	3.35	8.4	4.8(6.2^c)
(Cu1Co1)2Al-CLDHs	90	0.33	3.80	9.3	5.1
Cu2Al-CLDHs	61	0.26	4.02	4.7	15.4

---

3.3 BET

Fig. 3(I) shows the N₂ adsorption–desorption isotherms and the corresponding pore size distribution of (Cu_xCo_y)₂Al-CLDHs. All samples exhibit a typical IV isotherm with an H₃-type hysteresis loop, suggesting the presence of mesopores. This result is further confirmed by the corresponding pore size distribution in Fig. 3(II).

Fig. 3 N₂ adsorption–desorption curves (I) and pore size distributions (II) of (a) Cu₂Al-CLDHs, (b) (Cu₁Co₁)₂Al-CLDHs, (c) (Cu₁Co₂)₂Al-CLDHs, (d) (Cu₁Co₃)₂Al-CLDHs and (e) Co₂Al-CLDHs.

As can be seen from the pore size distributions, there are two main pore size distributions in the calcined catalysts. One is mesopores with sizes of several nanometers, and the other with a size of tens of nanometers. The former is proposed to be formed from the removed CO₃²⁻ and water in LDHs in the calcination process, and the latter came from the accumulation of oxide particles. From the SEM images (Fig. 4), it is seen that the particle size is in tens of nanometers, in accordance with the accumulation of oxide particles in the pores. A similar viewpoint has been stated in the literature.^28,45

For more detail, the pores formed from the remaining CO₃²⁻ and water can be divided into two categories. CO₃²⁻ ions were located between the sheets of LDHs, thus CO₃²⁻ leaving would cause slit-shaped pores between the sheets; while water in the state of hydroxyls was in the sheets of LDHs, thus water leaving would result micro and mesopores in the sheets. Also, due to the topological properties of LDHs, the calcined samples maintain the morphology, that is, the shape of the sheets were preserved. On the whole, the pores mainly came from the accumulation of particles, the accumulation of the oxide sheets formed the slit-shaped pores with several nanometer size, and the accumulation of the sheets constituted the larger particles in tens of nanometers which stacked and formed larger pores. This means that the surface area mainly came from the particles, hence no evident capillary condensation is exhibited in the sorption isotherms of Fig. 3.

The hierarchical porous structure plays a favorable role in the higher alcohol synthesis. Ding et al.⁴⁶ found that Cu–Fe supported on bimodal porous SiO₂ exhibited good catalytic activity and high selectivity to C₂₊OH, which was attributed to the well dispersion of active metal sites and high diffusion efficiency of products inside the bimodal porous structure. Lu et al.⁴⁷ found that three-dimensionally ordered macroporous Cu–Fe catalysts showed high activity and selectivity to C₂₊OH, because of the unique ordered porous structure of the catalysts, the uniformly distributed active components and the synergetic effect between Cu and χ-Fe₅C₂.

The specific surface areas and the pore structure parameters are shown in Table 1. The BET surface areas ranged between 61–137 m² g⁻¹ and the pore volumes between 0.26–0.47 cm³ g⁻¹. The specific surface area and pore volume decrease with the increase of copper content, which may be due to the different shape. With increasing cobalt content, the shape of particles changed from cubic or rectangular to platelet, see SEM images in Fig. 4a–c, and the platelets have a higher specific surface area. The shape difference in the morphology is likely due to the Jahn–Teller effect of Cu²⁺ ions, as pointed out in the literature.³¹

Fig. 4 SEM images of (a) Cu₂Al-LDHs, (b) (Cu₁Co₁)₂Al-LDHs, (c) Co₂Al-LDHs, (d) (Cu₁Co₂)₂Al-LDHs, (e) (Cu₁Co₂)₂Al-CLDHs and (f) (Cu₁Co₂)₂Al-Red.

3.4 SEM

The morphology of the samples was investigated by SEM as shown in Fig. 4. The LDH precursors display smooth and uniform nanocrystals with a narrow particle size distribution of 30–70 nm. Owing to the topological effect of hydrotalcite, when the LDHs were calcined and reduced, the same morphology of LDHs was still maintained, while the particle sizes became smaller than that of the precursors, see Fig. 4d–f. The similar morphology of the samples after calcination and reduction with LDHs confirmed the existence of the topological effect, supporting the above viewpoint that CuO, Co₃O₄ and CuCo₂O₄ are mixed evenly in the Al₂O₃ matrix, as a result, the formation of Cu–Co alloy is favored after reduction.

3.5 TEM

TEM pictures are presented in Fig. 5(a). A representative high resolution TEM image is shown in the inset picture in Fig. 5(a), and the lattice space of 2.08 Å corresponds to the (110) crystal plane of Cu–Co alloy, suggesting the formation of Cu–Co alloy, which is consistent with the XRD results in Fig. 1(III).

Fig. 5 (a) TEM images (insets: a HRTEM image and the particle size distribution diagram), (b) EDX mapping, (c) EDX spectrum of (Cu₁Co₂)₂Al-Red observed on a Mo substrate. Si comes from quartz sand. STEM images and the corresponding EDS line scanning profiles for Cu and Co elements of (d) and (d′) (Cu₁Co₁)₂Al-Red, (e) and (e′) (Cu₁Co₂)₂Al-Red, (f) and (f′) (Cu₁Co₃)₂Al-Red.

It can be seen from Fig. 5(a) that the Cu–Co alloy nanoparticles are evenly distributed throughout the amorphous Al₂O₃ matrix without severe aggregation and exhibit a uniform spherical-like shape. In addition, the average sizes of the Cu–Co alloy nanoparticles range from 3.5 to 7 nm, see the inset of particle size distribution in Fig. 5(a), which agrees with the average crystallite size calculated from the XRD data (listed in Table 1). The narrow particle size distribution is attributed to the uniform distribution of copper and cobalt ions in the precursor and to the effective prevention by the Al₂O₃ matrix of alloy particle sintering.

For further insight into the distribution of copper and cobalt, Energy Dispersive X-ray Spectrometry Mapping Analysis (Fig. 5(b)) was performed. Cu and Co show the same morphology of distribution, resulting in the uniform and homogeneous distribution of Cu and Co elements throughout the Al₂O₃ matrix. The energy dispersive X-ray spectrometry (EDX) analysis in Fig. 5(c) for (Cu₁Co₂)Al-Red reveals the co-existence of Cu, Co and Al elements, and the Cu/Co and (Cu + Co)/Al molar ratios are 0.49 and 2.08, respectively, which is in good accordance with that in the initial nitrate salt used for preparing the precursor.

Atomic-scale chemical composition across (Cu_xCo_y)₂Al-Red was investigated by HAADF-STEM-EDS and line scans, representative images and EDS curves are shown in Fig. 5d–f and d′–f′. The red lines in Fig. 5d–f in the STEM pictures are the scanning area, which corresponds to the EDS curves in Fig. 5d′–f′. The same variation tendency for the two elements of copper and cobalt can be seen clearly, resulting in the formation of homogeneous Cu–Co alloy in all the three catalysts. The intensity of the signal for copper and cobalt is different in different samples, indicating that the Cu–Co alloy formed with different Cu/Co ratios, which is in good accordance with the XRD and TPR results.

Studies on the phase diagram of binary Co–Cu suggest that the maximum solubility of copper in metal cobalt is 9% in molar ratio.⁴⁸ It should be noted that this solubility is for the bulk alloy, that is, for the alloy with large crystalline size. While for the nanosized bimetallic Cu–Co, the situation is different. Prieto et al.¹⁶ demonstrated that the superior nanoscale mixing of Cu and Co species can promote the formation of Cu–Co alloy nanocrystals after activation, and pointed out that the maximum occurrence of the Cu–Co alloy phase is at a Cu/Co molar ratio of 0.5. Volkova et al.¹⁸ also found single-phase Cu–Co alloy with the alloy composition Cu_0.5Co_0.5 after low-temperature reduction of CuCoO₂.

3.6 Catalytic performance

In the process of higher alcohol synthesis from syngas over Cu–Co-based catalysts, several parallel reactions including the Fischer–Tropsch synthesis and water–gas-shift reaction (WGSR) can occur. It is widely accepted that HAS requires both dissociative and non-dissociative activation of CO on different active sites. Cu can act as the active site for CO molecular activation and insertion, while Co for CO dissociative activation and chain propagation. An appropriate balance between CO dissociation and CO insertion has been suggested to be necessary for the synthesis of higher alcohols. So the synergetic effect of homogeneously dispersed copper and cobalt species plays a key role in the catalytic performance.

Fig. 6 shows the results of catalytic performance over the (Cu_xCo_y)₂Al-Red catalysts with different Cu/Co ratios. It is shown that the conversion of CO and the selectivity towards hydrocarbons increase with increasing cobalt content and reaction temperature over the three catalysts. This agrees with the fact that metallic Co⁰ is the active component for the F–T reaction, and is responsible for dissociative CO adsorption, C–C chain growth and hydrogenation. The selectivity of hydrocarbons increasing with reaction temperature is ascribed to the high activation energy for the dissociative adsorption of CO, which is favorable for hydrocarbon generation at higher temperatures.^3,15

Fig. 6 The conversions of CO and selectivity toward hydrocarbons, CO₂ and alcohols vs. reaction temperature over (■) (Cu₁Co₁)₂Al-Red, (●) (Cu₁Co₂)₂Al-Red and (▲) (Cu₁Co₃)₂Al-Red at a GHSV of 3900 mL (g_cat h)⁻¹ in the syngas mixture of H₂/CO/N₂ = 8/4/1 and at 3 MPa.

With the increase of copper content, the level of CO₂ production increased accordingly. This is in accordance with the fact that Cu is more active than Co for the WGSR.⁴⁹ In addition, an increased selectivity towards CO₂ with increasing reaction temperatures can be observed for all samples. The formation of alcohols and hydrocarbons are accompanied with the production of water, and water reacted with CO via the WGSR generated CO₂.⁵⁰ With the temperature increase, the content of water increased markedly with the increase of CO conversion, in turn accelerating the WGSR and CO₂ generation.

The catalytic performance results are presented in Table 2. The selectivity to higher alcohols increased from 42.3% with Cu/Co = 1 : 3 to 44.1% with Cu/Co = 1 : 2, and then went down to 42.6% with Cu/Co = 1 : 1. (Cu₁Co₂)₂Al-Red exhibited the highest selectivity toward higher alcohols, likely owing to the fact that most copper and cobalt were present in the alloy state, it is worth noting that the ratio of Cu/Co in (Cu₁Co₂)₂Al-Red is equal to that in CuCo₂O₄.

Table 2 Catalytic performance of CO hydrogenation over (Cu_xCo_y)₂Al-Red^a

Catalyst	XCO %	SRH %	SCO2 %	ROH distribution
				CH3OH	C2H5OH	C3H7OH	C4+OH
a Reaction conditions: 250 °C, GHSV = 3900 mL (gcat h)^−1, 3 MPa, H2/CO/N2 = 8 : 4 : 1.
Cu2Al-Red	35.7	11.8	8.5	76.2	2.1	0.9	0.5
(Cu1Co1)2Al-Red	43.2	50.1	5.3	2.0	18.0	15.6	9.0
(Cu1Co2)2Al-Red	51.8	51.3	2.9	1.7	18.3	16.9	8.9
(Cu1Co3)2Al-Red	54.0	53.5	2.4	1.8	18.5	14.6	9.2
Co2Al-Red	63.5	93.4	1.2	0.5	0.8	1.4	2.7

---

It is worth mentioning that the selectivity toward higher alcohols is higher than that over previously reported Cu–Co-based catalysts,^14,51 which is due to the formation of Cu–Co alloy and the uniform mixing of the elements in the catalyst. In the alloy phase, the two active sites are uniformly distributed and extremely close to each other. This can significantly increase the synergistic effect between copper and cobalt, which contributed to the enhancement of the selectivity to higher alcohols, as stated in the literature.^14,16,34

Over (Cu₁Co₂)₂Al-Red and at 250 °C, 3 MPa and a GHSV of 3900 mL (g_cat h)⁻¹, CO conversion is 51.8%, the selectivity to alcohols is 45.8%, and 94.3 wt% of the alcohols is C₂₊OH. The catalyst is very active and especially exhibited very high selectivity to C₂₊OH.

The stability test results for (Cu₁Co₂)₂Al-Red are shown in Fig. 7. The CO conversion and the selectivity to alcohols substantially remain stable in the 200 hour reaction process. After reaction for 200 h, the conversion of CO is 53.1%, the selectivity to alcohols is 38.4%, and the mass faction of higher alcohols in total alcohols is 92.8 wt%.

Fig. 7 Conversion of CO (■) and selectivity to alcohols (▼), hydrocarbons (●), CO₂ (▲) over (Cu₁Co₂)₂Al-Red with reaction time on stream at T = 260 °C, P = 3 MPa, GHSV = 3900 mL (g_cat h)⁻¹ and H₂/CO/N₂ = 8/4/1. The inset picture is the XRD profiles of (a) (Cu₁Co₂)₂Al-Red and (b) (Cu₁Co₂)₂Al-Red used for 200 h, (♦) Cu–Co alloy.

The inset picture in Fig. 7 is the XRD profiles of (a) (Cu₁Co₂)₂Al-Red and (b) (Cu₁Co₂)₂Al-Red used for 200 h. By comparing the diffraction peaks of Cu–Co alloy before and after the 200 h reaction, it is seen that the nanoparticles of Cu–Co alloy show good anti-sintering ability. The crystalline size only increased from about 4.8 nm for the catalyst before the reaction to about 6.2 nm for the catalyst after the stability test.

The catalytic performance of some representative excellent catalysts reported in literature is listed in Table 3. Taking activity and selectivity into consideration, Cu–Co alloy derived from (Cu_xCo_y)₂Al-LDHs turns out to be one of the best catalysts for HAS. In particular, compared with that of ref. 14, under the same reaction conditions, the as-prepared catalysts in our work show much better catalytic performance, for both the conversion of CO and the selectivity towards higher alcohols.

Table 3 CO conversions, alcohols selectivities and the corresponding reaction conditions for some representative catalysts reported in references

Catalysts	Temperature (°C)	H2/CO^a	Pressure (MPa)	GHSV (h^−1)	XCO (%)	SROH (%)	C2+OH^b%
a The molar ratio of H2/CO.b The mass fraction of C2+OH including ethanol in the total alcohols.c The unit is mL (gcat h)^−1.d In this work.e In this work.f In this work in stability test.
Co–Cu^49	240	1.5	4	7200^c	5.7	37.9	37
Co@Cu^42	230	2	2	18 000^c	─	33.0	91.4
CuFeMg^28	300	2	4	2000	56.89	49.1	77.0
Cu–Co/Al2O3 (ref. 14)	250	2	2	1800^c	23.2	23.3	79.3
Cu–Co/Al2O3^d	250	2	2	1800^c	47.6	38.2	88.1
Cu–Co/Al2O3^e	250	2	3	3900^c	51.8	45.8	94.3
Cu–Co/Al2O3^f	260	2	3	3900^c	53.1	38.4	92.8

---

4. Conclusions

Layered double hydroxides containing copper, cobalt and aluminum ions with a well crystallized hydrotalcite structure were prepared using a co-precipitation method. After calcination, the resultant oxide particles were uniformly mixed with a size of several nanometers and CuCo₂O₄ was one of the main phases, owing to the topological effect of LDHs. In the reduction process, the simultaneous reduction of copper and cobalt containing oxides and their uniform mixing resulted in the formation of Cu–Co alloy. XRD, TEM and SEM results confirmed or supported the formation of Cu–Co alloy. Also, the prepared catalysts possess dual pore size distribution structure with ranges of several nanometers and tens of nanometers.

The catalysts exhibited high activity and good stability for higher alcohol synthesis, especially showing very high selectivity to C₂₊ alcohols, which can be attributed to the formation of Cu–Co alloy and its high dispersion.

Acknowledgements

Financial support for this work was provided by the National Natural Science Foundation of China (NSFC) (nos 21376170, 21263011) is gratefully acknowledged.

Notes and references

1. R. Zhang, G. Wang and B. Wang, J. Catal., 2013, 305, 238.
2. J. Gong, H. Yue, Y. Zhao, S. Zhao, L. Zhao, J. Lv, S. Wang and X. Ma, J. Am. Chem. Soc., 2012, 134, 13922.
3. G. Liu, T. Niu, D. Pan, F. Liu and Y. Liu, Appl. Catal., A, 2014, 483, 10.
4. L. Shi, W. Chu and S. Deng, J. Nat. Gas Chem., 2011, 20, 48.
5. K. Xiao, Z. Bao, X. Qi, X. Wang, L. Zhong, K. Fang, M. Lin and Y. Sun, J. Mol. Catal. A: Chem., 2013, 378, 319–325.
6. E. Heracleous, E. T. Liakakou, A. A. Lappas and A. A. Lemonidou, Appl. Catal., A, 2013, 455, 145.
7. J. Hu, Y. Wang, C. Cao, D. C. Elliott, D. J. Stevens and J. F. White, Catal. Today, 2007, 120, 90.
8. S. F. Zaman and K. J. Smith, Catal. Today, 2011, 171, 266.
9. D. Li, C. Yang, N. Zhao, H. Qi, W. Li, Y. Sun and B. Zhong, Fuel Process. Technol., 2007, 88, 125.
10. H. Xiao, D. Li, W. Li and Y. Sun, Fuel Process. Technol., 2010, 91, 383.
11. K. Fang, D. Li, M. Lin, M. Xiang, W. Wei and Y. Sun, Catal. Today, 2009, 147, 133.
12. S. Deng, W. Chu, H. Xu, L. Shi and L. Huang, J. Nat. Gas Chem., 2007, 17, 369.
13. Y. Z. Fang, Y. Liu and L. H. Zhang, Appl. Catal., A, 2011, 397, 183.
14. J. Wang, P. A. Chernavskii, A. Y. Khodakov and Y. Wang, J. Catal., 2012, 286, 51–61.
15. G. Liu, Y. Geng, D. Pan, Y. Zhang, T. Niu and Y. Liu, Fuel Process. Technol., 2014, 128, 289.
16. G. Prieto, S. Beijer, M. L. Smith, M. He, Y. Au, Z. Wang, D. A. Bruce, K. P. de Jong, J. J. Spivey and P. E. de Jongh, Angew. Chem., Int. Ed., 2014, 53, 6397.
17. R. M. Bailliard-Letournel, A. J. Gomez Cobo, C. Mirodatos, M. Primet and J. A. Dalmon, Catal. Lett., 1989, 2, 149.
18. G. G. Volkova, T. M. Yurieva, L. M. Plyasova, M. I. Naumova and V. I. Zaikovskii, J. Mol. Catal. A: Chem., 2000, 158, 389.
19. X. Dong, X. L. Liang, H. Y. Li, G. D. Lin, P. Zhang and H. B. Zhang, Catal. Today, 2009, 147, 158.
20. C. J. Wang and D. O’Hare, J. Mater. Chem., 2012, 22, 23064.
21. J. L. Gunjakar, T. W. Kim, H. N. Kim, I. Y. Kim and S. J. Hwang, J. Am. Chem. Soc., 2011, 133, 14998.
22. P. J. Sideris, U. G. Nielsen, Z. Gan and C. P. Grey, Science, 2008, 321, 113.
23. G. V. Manohara, D. A. Kunz, P. V. Kamath, W. Milius and J. Breu, Langmuir, 2010, 26, 15586.
24. L. He, H. Berntsen, E. Ochoa-Fernández, J. C. Walmsley, E. A. Blekkan and D. Chen, Top. Catal., 2009, 52, 206.
25. A. I. Tsyganok, T. Tsunoda, S. Hamakawa, K. Suzuki, K. Takehira and T. Hayakawa, J. Catal., 2003, 213, 191.
26. K. Xiao, X. Qi, Z. Bao, X. Wang, L. Zhong, K. Fang, M. Lin and Y. Sun, Catal. Sci. Technol., 2013, 3, 1591.
27. C. Li, L. Wang, M. Wei, D. G. Evans and X. Duan, J. Mater. Chem., 2008, 18, 2666.
28. W. Gao, Y. Zhao, J. Liu, Q. Huang, S. He, C. Li, J. Zhao and M. Wei, Catal. Sci. Technol., 2013, 3, 1324.
29. S. Yang, L. Wang, S. Yue, Y. Lu, J. He and D. Zhao, Dalton Trans., 2014, 43, 8254.
30. A. I. Khan and D. O’Hare, J. Mater. Chem., 2002, 12, 3191.
31. H. Liu, Q. Jiao, Y. Zhao, H. Li, C. Sun and X. Li, J. Alloys Compd., 2010, 496, 317.
32. J. M. Oh, S. H. Hwang and J. H. Choy, Solid State Ionics, 2002, 151, 285.
33. X. Zhao, L. Wang, X. Xu, X. Lei, S. Xu and F. Zhang, AIChE J., 2012, 58, 573.
34. J. Su, W. Mao, X. C. Xu, Z. Yang, H. Li, J. Xu and Y. F. Han, AIChE J., 2014, 60, 1797.
35. A. R. Denton and N. W. Ashcroft, Phys. Rev. A, 1991, 43, 3161.
36. K. Takanabe, K. Nagaoka, K. Nariai and K. Aika, J. Catal., 2005, 232, 268.
37. F. Liu, Y. H. Qu, Y. Z. Yue, G. L. Liu and Y. Liu, RSC Adv., 2015, 5, 16837.
38. Ø. Borg, M. Rønning, S. Storsæter, W. van Beek and A. Holmen, Stud. Surf. Sci. Catal., 2007, 163, 255.
39. M. L. Smith, A. Campos and J. J. Spivey, Catal. Today, 2012, 182, 60.
40. A. Y. Khodakov, A. Griboval-Constant, R. Bechara and V. L. Zholobenko, J. Catal., 2002, 206, 230.
41. W. Chu, P. Chernavskii, L. Gengembre, G. Pankina, P. Fongarland and A. Khodakov, J. Catal., 2007, 252, 215.
42. N. D. Subramanian, G. Balaji, C. S. S. R. Kumar and J. J. Spivey, Catal. Today, 2009, 147, 100.
43. S. Varghese, M. G. Cutrufello, E. Rombi, C. Cannas, R. Monaci and I. Ferino, Appl. Catal., A, 2012, 443–444, 161.
44. X. Mo, Y.-T. Tsai, J. Gao, D. Mao and J. G. Goodwin, J. Catal., 2012, 285, 208.
45. S. Li, H. Wang, W. Li, X. Wu, W. Tang and Y. Chen, Appl. Catal., B, 2015, 166–167, 260.
46. M. Ding, J. Liu, Q. Zhang, N. Tsubaki, T. Wang and L. Ma, Catal. Commun., 2012, 28, 138.
47. Y. Lu, B. Cao, F. Yu, J. Liu, Z. Bao and J. Gao, ChemCatChem, 2014, 6, 473.
48. C. N. Ávila-Neto, D. Zanchet, C. E. Hori, R. U. Ribeiro and J. M. C. Bueno, J. Catal., 2013, 307, 222.
49. Y. Xiang, R. Barbosa and N. Kruse, ACS Catal., 2014, 4, 2792.
50. N. Tien-Thao, M. Hassan Zahedi-Niaki, H. Alamdari and S. Kaliaguine, J. Catal., 2007, 245, 348.
51. V. Mahdavi, M. H. Peyrovi, M. Islami and J. Y. Mehr, Appl. Catal., A, 2005, 281, 259.

---